The New Statesman’s rolling politics blog

RSS

The Tories' manipulation of education statistics

There is no evidence that reading standards have fallen among school children.

"Actual educational policy is being developed on the basis of false data." Photograph: Getty Images.

In Saturday's Guardian (Letters, 28 Jan), schools minister Nick Gibb defends the government's view that phonics are the only way to reach children to read. His central justification is that something must be done: "International studies rank England 25th for reading - down from seventh nine years ago."

In the very literal sense, Gibb is correct. In 2000, the OECD’s Programme for International Student Assessment (PISA) placed England in 7th position in its table (p.53). In 2009, it was in the 25th row of a similar table (p. 56).

In any other sense you care to mention, Gibb is entirely wrong , because:

1) Twelve other countries, nominally above England in the 2009 tables, have statistically insignificant higher scores. The National Foundation for Educational Research's summary of the OECD findings is quite explicit about this: "Because of the areas of uncertainty described above, interpretations of very small differences between two sets of results are often meaningless. Were they to be measured again, it could well be that the results would turn out the other way round (p.8)"

2) 31 countries took part in the tests in 2000, and 67 in 2009. Shanghai and Singapore may be nominally above the UK in the 2009 tables but they didn’t take part in the 2006 or earlier surveys. This makes direct comparison between years invalid.

3) The OECD’s warned explicitly (para 2 of this technical note) against comparing earlier PISA results with earlier data, because the very low response rate for earlier years largely invalidated samples.

4) The 2000 and 2003 tests were conducted some months earlier in school year 11 (Nov/Dec) than the 2006/2009 (March-May) ones, as an exception to the international study (to make room for GCSE preparations). As John Jerrim of the Institute of Education has noted, taking the tests around half a school year early makes a very obvious difference: "[I]t is important to understand that between November/December and March‐May of year 11 is likely to be a period when children add substantially to their knowledge of the PISA subjects as it is when pupils are working towards important national exams. Consequently, one should expect the year 11 pupils in the PISA 2000/2003 cohort to out‐perform their peer taking the test in 2006/2009 due to the extra five months they have had at school….."

In short, there is simply no reliable evidence that 15-year-olds in England are any less able to read and understand texts, when compared to their international peers, than they were nine year ago. Yet here we have a government minister using that argument as a key reason for a fundamental and controversial change in which five and six-year-olds are taught.

Now, if this was a result of incompetence on the part of the minister and his department, that would be worrying enough. But what should really concern us is that the Department of Education almost certainly knows perfectly well that its "interpretation" of the OECD data is entirely incorrect, but is determined to carry on peddling its untruths anyway.

The key evidence of this, I suggest, is the way in which Michael Gove himself defended his proposals for a return to 'O' Levels/CSE in parliament on 21 June:

The sad truth is that, if we look at the objective measure of how we have done over the past 15 years, we find that on international league tables our schools fell in reading from 523 to 494 points, in maths from 529 to 492 and in science from 528 to 514.

Here, Gove used the OECD raw scores for 2000 and 2009 rather than the table rankings (the lower scores can largely still be explained by two of the factors above). He almost certainly did this because he and his team realised they had been rumbled by blogs like Though Cowards Flinch with a mind to detail, and by a Guardian editorial of the same day, which said:

Mr Gove.... latches on to data purporting to show English schools plummeting down world rankings. The Institute of Education has meticulously documented all sorts of distortions in these apparently alarming figures, but such calming analysis fails to register. Mr Gove should go away, revise the evidence properly – and prepare for a resit.

Clearly, Gove didn't want to be caught red-handed by Labour members assiduous enough to have read the Guardian that morning. Yet just a month later we have the schools minister writing to the same paper with the very nonsense his boss had been wary of using.

The real tragedy, of course, is not that Guardian readers are being lied to, but that actual educational policy is being developed on the basis of false data. The direct consequence of the pretence that comparative reading standards are plummeting is a emphasis on setting higher targets, as set out by Sir Michael Wilshaw, the Chief Inspector of Schools who, sadly, has been all too complicit in the myth-making. Wilshaw has stated that: "So one of the first questions we need to ask is whether the national end-of-primary-school target of level 4 is sufficiently high to provide an adequate foundation for success at secondary school."

Yet data in the government's own 2010 education white paper suggests that the actual problem policymakers should be facing up to is not low targets, but unequal distributi on of achievement between the upper and lower percentiles compared with other countries (see Exhibit 1.1 in this PIRLS report). By focusing their energies on the creation of fundamentally dishonest headlines, the government and its advisers are actively missing out on data which might actually improve the lives of young people.

Of course, this is not the first time that the government has resorted to the use of dodgy statistics. Chris Grayling has already had his wrists slapped by the UK Statistics Authority for his flagrant abuse of statistics. Now, it even looks as though the government may attemp to explain away its disastrous management of the economy by casting doubt on the reliability of the GDP data collected by the Office for National Statistics, without providing a shred of evidence as to how these dataset might have been considered reliable for so long but are now, so suddenly, suspect.

Overall, a picture is starting to emerge of a government prepared, in its mix of desperation and ideological fervour, to go one step beyond spin. That should keep us on our toes.